Acc. Chem. Res. 1988,21, 307-312
307
Elemental Fluorine as a “Legitimate”Reagent for Selective Fluorination of Organic Compoundst SHLOMO ROZEN School of Chemistry, Raymond a n d Beverly Sackler Faculty of Exact Sciences, Tel-Auiv University, Tel-Auiu 69978, Israel Received J a n u a r y 19, 1988 (Revised Manuscript Received May 10, 1988)
In a sense, elemental fluorine is one of the oldest potential reagents in chemistry. Its isolation more than a century ago resulted in the award of the Nobel Prize to the French chemist Moissan in 1906.l Moissan himself started experiments on organic substrates with fluorine as a reagent? but the results were such that the next set of experiments was conducted only a generation later, this time with the important finding that diluting the fluorine with an inert gas and lowering the reaction temperature resulted in some interesting nonviolent reactions mainly of radical naturea3 However, these results did not attract many followers, apparently because fluoro-organic chemistry was still in its infancy and because chemists were too well aware of the extreme reactivity of the first halogen. The Second World War generated a very intense interest in fluorine chemistry, mainly associated with the Manhattan Project. This required chemicals with extreme properties such as uranium derivatives with high volatility and materials with high stability toward very corrosive conditions. The former requirement was satisfied by compounds such as UF6, while the latter was found in fluorine-containing polymers. Although interesting biological activity had already been discerned in certain fluoro-organic derivatives before the war, the boom in this area started mainly in the 1950s, along with an accelerated development of fluorinated solvents and polymers. Today many billions of dollars are associated with fluoro-organic chemistry, thousands of relevant publications appear every year, and many chemists devote their entire research to this field. It is therefore somewhat surprising to notice that until recently most if not all of this chemistry was based on nucleophilic reactions using various types of fluorides. A few reactions were also accomplished by various radical pathways, again using fluorine salts such as CoF3 and HgF2. One notable exception was perchloryl fluoride (FC1Os),which tended to attack electron-rich sites associated with very activated double bondsa4 Its common use, however, was relatively short, since a number of tragic explosions occurred whose origin was the frequently developed chlorate m ~ i e t y . ~ Professor Barton was the first to introduce the term electrophilic fluorination. He employed the known, but new to organic chemistry, CF30F.6 Shortly afterward, we7 and many others8 started to work with this comProfessor Shlomo Rozen was born in 1942 in Bulgaria. He received his B.Sc., M.Sc., and Ph.D. at the Hebrew University In Jerusalem, where he worked with I. Shahak and the late Ernst D. Bergmann, the pioneer of fluorine chemistry in Israel. He then spent 3 years at the Research Institute for Chemistry and Medicine in Cambridge, MA, with D. H. R. Barton, R. H. Hesse, and M. M. Pechet, where h e began his work wlth elemental fluorlne. In 1976 h e joined the School of Chemistry at Tel-Aviv University. Since then h e has also spent nearly two years in the Central Research Department of E. I. du Pont d e Nemours in Wilmington, DE.
mercially available gas in which the oxygen-bound fluorine acts as an electrophile. Unfortunately, after a few years this new fluorinating agent ceased to be commercially available, and the number of the relevant publications dropped drastically. The O-F bond, however, greatly resembles the bond in the fluorine molecule (42 vs 39 kcal/mol), a fact that encouraged us to evaluate the potential of fluorine itself in preparative organic chemistry. At the beginning of the 1970s one very important use of F2 began to emerge. Lagow showed that when properly handled, fluorine could replace all the hydrogens in a given organic molecule by stepwise radical reactions, thus affording important perfluoro derivat i v e ~ . Our ~ interest, though, was aimed mainly at selective monofluorination, since many compounds of this type have a pronounced biological activity. The main question was whether regio- or even stereoselective reactions could be achieved, since expecting selectivity from this most reactive of elements seemed at first to be almost a contradiction in terms.1°
Direct Fluorination A. Electrophilic Activation of C-H Bonds. Electrophilic reactions on saturated carbon-hydrogen bonds are known and comprehensively summarized in an excellent recently published book by 0lah.ll Still, the existing examples are very few in comparison with general nucleophilic reactions or electrophilic ones on Dedicated to Professor D. H. R. Barton on the occasion of his 70th birthday. (1) In 1986, during the events dedicated to the centenary of the discovery of fluorine, the French Post Office honored Moissan for his achievements by issuing a beautiful commemorative stamp. In this stamp, however, the obvious reaction of F, H2 2HF was presented instead of the one describing Moissan’s main achievement: 2HF F2 H,. (i)Moissan, H. Ann. Chim. Phys. 1891, 19, 272. (3) Bockemuller, W. Justus Liebigs Ann. Chem. 1933,506,20. Miller, W. T.;Calfee, J. D.; Bigelow, L. A. J. Am. Chem. SOC. 1937, 59, 2072. Fukuhara, N.: Bigelow,-L. A. J. Am. Chem. SOC. 1938, 60, 427. (4) Djerassi, C.; Chamberline, J. W. Steroid Reactions; Holden Day: San Francisco, 1963; p 155. (5) Sharts, C. M.; Sheppard, W. A. Org. React. 1974, 21, 125. (6) Barton, D. H. R.; Godhino, L. S.; Hesse, R. H.; Pechet, M. M. J . Chem. SOC., Chem. Commun. 1968,804. Barton, D. H. R.; Danks, L. J.; Ganguly, A. K.; Hesse, R. H.; Tarzia, G.; Pechet, M. M. J. Chem. SOC., Chem. Commun. 1969, 227. (7) Rozen, S.; Shahak, I.; Bergmann, E. D. J. Org. Chem. 1975, 40,
+
+
-
-
~~
2966. (8) Hesse, R. H. Isr. J. Chem. 1978, 17, 60. (9) Adcock, J. L.; Lagow, R. J. J. Am. Chem. SOC. 1974, 96, 7588.
Many useful variations to this method have been introduced since. See, for example: Lin, W. H.; Bailey, W. I., Jr.; Lagow, R. J. J. Chem. SOC., Chem. Commun.1985,1350. Persico, D. F.; Gerhardt, G. E.; Lagow, R. J. J. Am. Chem. SOC. 1985,107,1197. Adcock, J. L.; Cherry, M. L. Ind. Eng. Chem. Res. 1987,26,208. Adcock, J. L.; Robin, M. L. J.Org. Chem. 1984, 49, 1442 and references therein. (10) Recently, the link between reactivity and selectivity has been reevaluated Pross, A. Isr. J. Chem. 1986,26,390 and references therein. (11) Olah,G. A.; Pra@sh, G. K. S.; Williams, R. E.; Field, L. D.; Wade, K. Hypercarbon Chemtstry; Wiley: New York, 1987.
0001-4842/88/0121-0307$01.50/0 0 1988 American Chemical Society
Rozen
308
unsaturated centers. If fluorine were induced to react in an ionic mode, it would of course be a very strong electrophile and would have the advantage over others in its efficiency and in the ease of product identification. One factor that may suppress rival radical reactions is a low fluorine concentration. This, however, was not enough, since when fluorine, diluted by an inert gas, was passed through solutions, in a nonpolar solvent, of suitably protected 4-methylcyclohexanol, only fluorine-containing tars were produced. The reaction course was altered dramatically by a change in the solvent system. When the same reaction was repeated, this time in a 1:l mixture of CFC13and CHCl,, substitution of the tertiary hydrogen at C-4 by fluorine occurred in 60% yield. This proceeded with full retention of configuration in both the cis and the trans series (eq 1).l2
Accounts of Chemical Research
tected 2-methylbutanol and nopinone, where the degree of hybridization of the tertiary C-H bond is very low because of either the proximity of the oxygen or the involved strain. In neither case was any specific substitution observed, while with protected 6-methyl-2heptanol or adamantane derivatives the yields of specifically fluorinated products were very good. The uniqueness of this substitution reaction lies in the activation of sites that usually are extremely unreactive to other reagents. Such fluorination provides an excellent entry for further chemical transformations. This can be accomplished via dehydrofluorination, using BF3, HF, NaOH, or MeMgX (eq 3).13
A
R&r"R HwoR amoR + HF
tZ
-
[RH.FAF
Hcc,]-+RF
+ cls(F).
trans(HI+
H
~ 1 6 ( H )
(1)
a
(')
t
tl
li
no reactlo".
X
CIB
and 1ran6
X
X - H
= F
The same trend was found in both the aliphatic15and the bicyclic series.16 Examples are presented by pro-
s.;Gal, C.; Faust, Y. J.A m . Chem. SOC.1980,102, 6860. (13) Rozen, S.; Gal, C. J. Org. Chem. 1987, 52, 2769. (14) Rozen, S.; Gal, C. Tetrahedron Lett. 1984,25, 449. (15) Rozen, S.; Gal, C. J. Org. Chem. 1987, 52, 4928.
(12) Rozen,
-HF X = H-%X
= F
1
It was intriguing to use this route to try to prepare the yet unknown tetra-tert-butylethene (eq 4). The
trans(F)
The addition of chloroform, or other polar solvent, such as AcOH or CH3NOZ,serves several purposes: it is a good radical scavenger; it provides a polar medium, which is helpful whenever an ionic type reaction is to be encouraged; and not least, it acts as an acceptor for the developing F- through hydrogen bonding. The substitution mechanism that accommodates all the observations is one involving the nonclassical carbocation A (eq 1)resulting from the attack of the positive pole of the fluorine molecule on the electrons of the tertiary C-H bond.13 Such a mechanism, typical of electrophilic reactions," should lead to a full retention of configuration, but this is the first time that such a retention has clearly been shown to exist in all the examples studied. This type of electrophilic u substitution bears some similarities to the more familiar process around regions rich in T electrons. It is thus not surprising that the reaction is largely governed by electronic factors and will target mainly on tertiary C-H bonds. Indeed, calculating the hybridization of all the hydrogens in many molecules revealed that the attack always took place on the C-H bond with the highest p-orbital cont r i b ~ t i 0 n . lIt~ is thus understandable that a substitution of a tertiary hydrogen with an exceptionally low hybridization, such as one on a cyclopropane ring or in a very close proximity to an electron-withdrawinggroup, could not take place, and the starting materials were recovered even after prolonged treatment with Fz. On the other hand, when the electron-withdrawing group is more distant from the tertiary center, or when none is present as in paraffins,14 the yields increased considerably (eq 2). Me-&CHZCWR,
X = H % X = F
OAc
0AC
(41
attempt, however, failed since the fluorine molecule has to approach the tertiary C-H bond with a t least one molecule of solvent. The steric factor, which usually is not of dominant importance, proved in this extreme case to be insurmountable. Fluorination of steroids demonstrates some subtle electronic and steric effects.17 While four tertiary hydrogens are present in 5a-androstane-3/3,17/3-dioldiacetate, only the remotest 9a-H was replaced by fluorine.18 However, it is possible to substitute other tertiary hydrogens by attaching various deactivating groups on different sites of the steroidal ~ke1eton.l~In each case subsequent dehydrofluorination provides a good route for further chemical transformations20 on sites that were previously completely inert (eq 5).19v21
c
(16) Rozen, S.; Gal, C. Tetrahedron Lett. 1985, 26, 2793. (17) Actually we started the research on steroidal tertiary fluorination using CF,OF and Fz while working a t the Research Institute of Medicine and Chemistry in Cambridge, MA, under Profs. Barton, Hesse, and Pechet about 15 years ago. I would like to use this opportunity to thank them again for a great time. (18) Barton, D. H.R.; Hesse, R. H.; Markwell, R. E.; Pechet, M. M.; Rozen, S. J. Am. Chem. SOC.1976, 98, 3036. (19) Alker, D.; Barton, D. H.R.; Hesse, R. H.; James, J. L.; Markwell, R. E.; Pechet, M. M.; Rozen, S.; Takashita, T.; Toh, H. T. Nouu. J.Chim. 1980, 4, 239. (20) The synthesis of many steroidal drugs starts from the corresponding olefins. A As,11steroid provides a starting point for cortisone type drugs possessing the 9oc-fluoro-ll(3-hydroxymoiety. A double bond at 14,15 can serve as a precursor for many cardenolides, most having a hydroxyl at 14, while cholesterol with a double bond at 17,20 is potentially an inexpensive starting material for the synthesis of many other expensive steroids.
Fluorine as a Reagent for Selective Fluorination
Vol. 21, 1988
The less common Sa-acetoxy steroids allowed fluorination of the nearby 5-position (never observed in the 36 series), apparently by stabilizing the appropriate transition state.22 With the 56 steroids the cis A/B arrangement prevented the fluorine from approaching the 9a-H (although electronically favored) and the fluorination took place always on the next most suitable tertiary site (eq 6).22 These rigid and well-defined
AcO
309
fluorine to certain simple olefins, but his technique was quite unusual and inconvenient. Many years later fluorine was bubbled through a solution containing a steroidal enone, which resulted in the formation of the corresponding a,$difluoro derivative. This was accompanied by a rearrangement product (eq 8),29show\
\
\
c=o
c=o
c=o
& AcodP
X = Y = H -
(6)
X = H ; Y = F
+
X = P ; Y = H
fluoro steroids presented a good opportunity for detailed 13C NMR studies. From these it was concluded that in such systems all y carbons in anti position were deshielded and frequently coupled by the fluorine atom, while the gauche y carbons were shifted to higher field by up to 5 B. Reactions with Organic Halides. Fluorine when acting as an electrophile can attack the electrons of other elements in organic substances. No easy attack is observed on oxygen, chlorine, and fluorine atoms since their bonds with carbon are quite strong, but it can attach itself to the nitrogen lone pair, resulting in some cases in destruction of the original compound24 while in others in the formation of new and useful fluorinating agents.25 F2also reacts immediately with aliphatic bromides and iodides, which then decompose to the corresponding halogen fluorides and many undefined products. However, in cases where the resulting carbocations are highly stabilized as in adamantane derivatives, fluorine after an initial electrophilic attack on the halogen will eventually substitute it in excellent yields, forming the corresponding fluoroadamantanes. If sources of other nucleophiles such as C1- or RO- are present, chloro- or alkoxyadamantane derivatives are also obtained (eq 7).26
C. Addition to Double Bonds. In contrast to the direct chlorination or bromination of olefins, the preparation of 1,2-difluoro derivatives was considered to be difficult, and ways of circumventing the use of the element itself have been developed.27 The pioneering work of Merritt2s has shown that it is possible to add (21) It should be mentioned that the elegant radical “remote control” activation developed by Breslow has also led to tertiary halogenation and eventually to similar steroidal olefins; see, for example: Breslow, R. Acc. Chem. Res. 1980,13, 170. (22) Rozen, S.;Ben-Shoshan, G. J. Org. Chem. 1986,51, 3522. (23) Rozen, S.; Ben-Shoshan, G. Org. Mugn. Reson. 1985, 23, 116. (24) Rozen, S.;Gal, C., unpublished results. (25) Purrington, S. T.; Jones, W. A. J . Org. Chem. 1983, 48, 761. Barnette, W. E. J.Am. Chem. SOC.1984, 106, 452. Umemoto, T.; Kawada, K.; Tomita, K. Tetrahedron Lett. 1986,27,4465. (26) Rozen, S.;Brand, M. J. Org. Chem. 1981,46, 733. (27) See, for example: Bornstein, J.; Borden, M. R.; Nunes, F.; Tarlin, H. I. J. Am. Chem. SOC.1963,85,1609. Sket, B.; Zupan, M. J. Chem. SOC.,Perkin Trans. 1 1977, 2169. Shellhamer, D. F.; Conner, R. J.; Richardson, R. E.; Heasley, V. L. J. Org. Chem. 1984,49, 5015. (28) Merritt, R. F.; Johnson, F. A. J. Org. Chem. 1966, 31, 1859. Merritt, R. F.; Steven, T. E. J.Am. Chem. SOC.1966,88, 1822. Merritt, R. F. J. Org. Chem. 1966,31,3871. Merritt, R. F. J.Am. Chem. SOC.1967, 89, 609.
ing that the addition is not a four-centered process, suggested previously as a possible mechanism.28 Later we found that adding a proton donor to the solution (e.g., EtOH) helped produce the expected difluoro adduct in a cleaner and better yield reaction.30 The stereoselectivity is usually very high and, contrary to other halogens, was found to be syn, characteristic of an addition process involving electrophilic fluorine (eq 9).31 Similar additions were also obtained with some sugars32and heterocycle^,^^ enabling the positron-emitting radioisotope 18F to be introduced. EICH=CH(CHz)zOAe
trans
5
1
E t C H - C H ( C H 2 ) 2 0 A c d EtCH-CH(CHZ)20Ac
[A:t i g h t ion p a i r
!A
(9)
threo erythro
01s
Fluorine was successfully added to the enolic forms forming the corof pyruvic acid derivatives (eq RCHzCOCOOR‘
RCHPCOCOOR‘ ; AcO
n]
-% many undefined
Compounds
(10)
responding fluoropyruvates. Addition to enols, however, did not prove to be a general reaction and when Fzwas passed through an electron-rich enol acetate, an instantaneous reaction took place with complete decomposition of the substrate. Is fluorine then limited in its reactions to tertiary H substitution and addition to a few olefins? The answer is definitely no, although somewhat indirect methods have to be used.
Indirect Fluorination A. Using Perfluorinated Fluoroxy Compounds. Several compounds possessing the weak 0-F bond are known. They can be grouped into two categories: one with the OF moiety bonded directly to some perfluorinated residue such as Rf or SF5and the other to a perfluoroacyl group-RfCOOF. CF30F, the only compound produced commercially, received the most attention,s but its present price and availability have discouraged many from working with it. Most other fluoroxy compounds have been prepared mainly by inorganic chemists in order to measure physical constants, rather than for organic synthesis.35 CF3COOF,which (29) Barton, D. H. R.; James, J. L.; Hesse, R. H.; Pechet, M. M.; Rozen,
S. J. Chem. SOC.,Perkin Trans. I 1982, 1105. (30) Rozen, S.;Brand, M. J. Org. Chem. 1986,51, 3607. (31) Barton, D. H. R.; Hesse, R. H.; Jackmann, G. P.; Ogunkoya, L.; Pechet, M. M. J . Chem. SOC.,Perkin Trans. I 1974, 739. (32) Ido, T.; Wan, C. N.; Fowler, J. S.;Wolf, A. P. J. Org. Chem. 1977, 42, 2341. Diksic, M.; Jolly, D. Carbohydr. Res. 1986, 153, 17. (33) Shiue, C. Y.; Wolf, A. P. J.Labelled Compd Radiopharm. 1980, 18, 1059. (34) Tsushima, T.; Kawada, K.; Tsuji, T.; Misaki, S. J. Org. Chem. 1982, 47, 1107. (35) Most of this work had been performed by Prager and Thompson in the mid 1960s. See, for example: Prager, J. H.; Thompson, P. G. J. Am. Chem. SOC.1965,87, 230. Thompson, P. G.; Prager, J. H. J. Am. Chem. SOC.1967,89, 2263. For a relevant review, see: Mukhametshin, F. M. Usp. Khim. 1980,49, 1260.
Rozen
310
was prepared in a unique and difficult way,%is a typical example. Years later we passed dilute fluorine through a suspension of CF3COONa in CFC13 and obtained an oxidizing solution.37 The oxidative power was due to a mixture of several compounds, but a feature common to them all-the OF moiety-enabled us to regard it as a single homogeneous reagent. Thus, without any tedious isolation and purification steps, it was reacted with many types of electron-rich enol acetates, forming eventually the corresponding a-fluorocarbonyl derivatives (eq ll).38It should be noted that the reactions
Accounts of Chemical Research
elsewhere. First, the oxidizing solution was found to be stable at 0 “C for at least a year, thus serving as a “shelf reagent” for electrophilic fluorinations. Second, and probably more important, is the fact that the higher hypofluorites are excellent radical initiators for the polymerization of tetrafluoroethene and other fluorine-containing olefinic monomers. Their unique advantage over other initiators is that no harmful end groups are formed, since the initiator decomposes to C02and Rf‘, the latter becoming an integral part of the polymer (eq 13).43 CF3(CF2)6COOK
+
F2
+ CPQ(CFZ),OF
are technically simple and the yields comparable to those obtained with the expensive CF30F,proving that after all, elemental fluorine is capable of reacting reasonably well with enols, if indirectly. We found that under anhydrous conditions a large proportion of the above oxidizing mixture was CF3CF20F, while in the presence of water or H F mostly CF3COOFwas formed. This presented an opportunity to introduce the perfluoroethoxy moiety into various systems39and increase their lipophilicity, an important factor in many potential drugs (eq 12). An interesting modification of the preparation of fluoroxypentafluoroethane containing lSF has recently been publi~hed.~O PhCHsCKR
*-F 2
+
CF3COONa
PhCHd-R
1 PhCH-CHR
I
F
I
I
RC=CHR’
+
R
I
OH
Similarly, trifluoroacetyl hypofluorite (CF3COOF) was added to alkenes, forming the corresponding fluorohydrin (eq 12),4lan important moiety in many drugs. The reaction of both CF3CF20Fand CF3COOF is fully regiospecific and has good stereoselectivity. The synmode addition, characteristic of electrophilic fluorination is dominant, with CF,COOF giving better stereoselectivity than CF30F42or CF3CF20F.39In both cases when the carbocation in the intermediate tight ion pair (eq 9) is stabilized more than usual, some of the stereoselectivity is lost, proving again that a four-centered reaction is not involved. Fluorine can also react with salts of higher perfluorinated acids. The products are fluoroxy compounds of the type R@F and RfCOOF, depending on the condition^.^^ The oxidizing components (either a mixture or a relatively pure species) are able to fluorinate enol acetates (eq ll),but their importance lies (36)Gard, G.L.; Cady, G. H. Inorg. Chem. 1965,4,594. (37)Rozen, S.;Lerman, 0. J . Am. Chem. SOC.1979, 101,2782. (38)Rozen, S.;Menahem, Y. J. Fluorine Chem. 1980,16,19. For a review on the synthesis of a-fluorocarbonyl derivatives, see: Rozen, S.; Filler, R. Tetrahedron Rep. 1985,41, 1111. (39)Lerman, 0.; Rozen, S. J . Org. Chem. 1980,45,4122. (40)Mulholland, G.K.; Ehrenkaufer, R. E. J . Org. Chem. 1986,51, 1482. (41)Rozen, S.;Lerman, 0. J . Org. Chem. 1980,45, 672. (42)Barton, D.H. R.; Hesse, R. H.; Jackmann, G. P.; Ogunkoya, L.; Pechet, M. M. J. Chem. Sot., Perkin Trans. 1 1974,739. (43)Barnette, W. E.;Wheland, R. C.; Middleton, W. J.; Rozen, S. J . Org. Chem. 1985,50, 3698.
Rf
:c=cAcOF
:C-C:
I I
F
(12) ___*
RftCF2-CFZ
Up to that time all compounds with the OF moiety had a perfluoro residue. It was assumed that if this reactive moiety were bonded to an alkyl radical, an immediate HF elimination would take place, as theoretically predicted for the nonexistent CH30F.44 As happened to many legends connected with F2,this was also found to be of limited value. We were able to show that in acetyl hypofluorite, elimination is hindered by the interposed spacer between F and the adjacent alkyl hydrogens. B. Acetyl Hypofluorite. When we passed dilute fluorine through a suspension of sodium acetate in CFC13, an oxidizing solution was obtained that reacted with olefins, adding the elements of AcOF (eq 14).45
I
&XzCF3
F2 + CF3COONa/HZ0 CFQCOOF
NaOH
PhCH-CHR
CF3CF20F
RfCOOF
+
C F ~ ( C P Z ) ~ C F ( O P )+~ CF3(CF2)6COO~
0
OAc
-(OF)
+
A C O ~ ~t F AcO
(14)
OAC
K & r
HOCH,
no
“F
Since there was a possibility that the reactions of this oxidizing solution did not originate from a single reagent, but rather from two or more consecutive steps providing the elements of AcO and F,46some of the molecule’s physical constants and thermal characteristics were studied.47 Eventually AcOF, the first member of a new oxidative family of acyl hypofluorites, was isolated in pure form by Appelman, proving its existence beyond any It became clear that the advantage of acetyl hypofluorite lies in its relative mildness. While the other fluoroxy reagents react cleanly only with benzylic or electron-rich double bonds, AcOF gives good results with most types of olefins.49 Since the reactions are fast and proceed with good yields and syn stereospecificity, they were immediately mobilized also for the incorporation of the 18F into biologically interesting compounds. For example, 2-deo~y-2-(~~F)fluoro-~glucose was prepared (eq 14) and used for extensive brain studies by using positron emission transaxial to(44)Wright, J. S.; Salem, L. J . Am. Chem. SOC.1972,94,2371. (45)Rozen, S.;Lerman, 0.; Kol, M. J. Chem. Sot., Chem. Commun. 1981,443. (46)A similar case was observed when the elements of CH30 and F were added to olefins (see ref 42). (47)Hebel, D.; Lerman, 0.;Rozen, S. J.Fluorine Chem. 1985,30,141. (48)Appelman, E.H.; Mendelsohn, M. H.; Kim, H. J. Am. Chem. Sot. 1985,107,6515. (49)Rozen, S.;Lerman, 0.;Kol, M.; Hebel, D. J . Org. Chem. 1985,50, 4753.
Fluorine as a Reagent for Selective Fluorination
Vol. 21, 1988
mography (PETT).50 Because of the importance of AcOF, especially as an 18Fcarrier, several variations for its preparation have been developed, the main one involving the passage of F2 through a column packed with various salts of acetic acid.51 The mildness of AcOF is evident from its smooth reaction with 1,3-dicarbonyl compounds or their enolate^.^^ What is more, even lithium enolates of monocarbonyl derivatives could be directly flu~rinated?~ eliminating the need to prepare the corresponding enol acetates, silyl enol ethers, and the like. The parallel reactions with F2, CF30F, or CF3COOF resulted only in tars or in very low yields of the expected monofluoro derivatives (eq 15). o Li' +
xF+
; RC-CR'2
X = CHICOO:
929
X = CF3COO:
13%
X
= F:
+ AcOF
+
RCOCFRVZ
0%
311
bonds forming the corresponding ArF compound^.^^ After some of its fundamental reactions had been studied, only a short period elapsed before efforts were channeled toward using AcOF for electrophilic fluorinations of compounds with biological interest. Thus several fluoro sugars have been prepared,%the reactions of AcOF with uracil and cytosine were thoroughly and fluoroantipyrine@' and fluorodopa61were obtained. Most of these compounds have been made with the 18Fradioisotope and hence on a small scale. Since we do not have the facilities for making and studying 18F-containingcompounds, we have restricted ourselves to the development of the chemical processes and to demonstrating that it is possible to make in good yields and on a larger scale, many types of biologically interesting derivatives, including fluorohexestrols, fluorotyrosine, various 4-flUOrO steroidal enones, 6fluoro and fluorobimans, which are important fluorescent compounds with useful applications in biology (eq 17).63
COOMe Though relatively mild in action, in absolute terms AcOF is a powerful reagent. One of the reactions emphasizing this fact is aromatic fluorination. The inF corporation of a fluorine atom into a benzenoidic ring OMe F is achieved routinely only by the time-honored BalzX = H ; P (11) Schiemann reaction starting from the appropriate anThe free radical nature of fluorine reactions has been iline derivative. There have been several attempts to long known. We have demonstrated that the same fluorinate aromatic compounds with F2 or fluoroxy element can also be a source for electrophilic fluorine. compounds, but the conversions had to be kept unpracticably low if yields were not to drop d r a ~ t i c a l l y . ~ ~ The versatility of this halogen is revealed in full by considering some new reactions in which this element AcOF, however, could be reacted with many activated can be an immediate source for nucleophilic fluorine as aromatic compounds to produce mainly the ortho fluoro well, again without isolating or purifying the intermederivatives in very good conversions and yields of up diate reagents. to 85%.55 Several explanations have been offered for C. Iodine a n d Bromine Monofluorides. Adding the dominant ortho substitution in similar reactions, the elements of halogen fluorides across a simple double including cyclic intermediates (for the reaction of PhOH bond is a well-known procedure, achieved by a one-pot, and Ft6), but we have shown that AcOF is sufficiently two-step reaction using HF and an N-halo amide. This powerful to add across the most electron-rich region of procedure, however, does not employ the actual halogen the aromatic ring in a fast reaction, suitable also for fluoride molecule, imposing quite a few limitations on working with 18F. A subsequent spontaneous eliminathe scope of the reaction. Fluorine, being an exception of AcOH restored the aromaticity, but in cases tionally active element, reacts with other halogens, where this was not possible, the resulting cycloproducing the corresponding XF molecules. Among hexadiene reacted very rapidly with the reagent and them ClF is a stable compound and so its chemistry has tars were obtained. With careful monitoring, however, to the been considerably explored.64 This is in contrast the corresponding adducts could in some cases be isoother two monofluoro interhalogens. IF has been delated (eq 16).55 Acetyl hypofluorite, mainly with 18F, scribed in only a few publications, most of them dealing was also used for cleavage of Ar-Sn, Ar-Si, and Ar-Hg with its thermal stability in the gas phase65 and its
X = OR, NHCOR
R = H ; CHO
(50) Shiue, C. Y.; Salvadori, P. A.; Wolf, A. P. J. Nucl. Med. 1982,23, 108. Adam, M. J. J. Chem. SOC.,Chem. Commun. 1982, 730. Ehrenkaufer, R. E.; Potocki, J. F.; Jewett, D. M. J. Nucl. Med. 1984,25, 333. (51) Jewett, D. M.; Potocki, J. F.; Ehrenkaufer, R. E. Synth. Commun. 1984,14,45. Jewett, D. M.; Potocki, J. F.; Ehrenkaufer, R. E. J. Fluorine Chem. 1984,24,411. (52) Lerman, 0.;Rozen, S. J. Org. Chem. 1983, 48, 724. (53) Rozen, S.; Brand, M. Synthesis 1985, 665. (54) Cacace, F.; Giacomello, P.; Wolf, A. P. J . Am. Chem. Soc. 1980, 102, 3511. Fifolt, M. J.; Olczak, R. T.; Mundhenke, R. F.; Bieron, J. F. J . Org. Chem. 1985, 50, 4576. (55) Lerman, 0.;Tor, Y.; Hebel, D.; Rozen, S. J. Org. Chem. 1984,49, 806. (56) Misaki, S. J. Fluorine Chem. 1981, 17, 159.
(57) Adam, M. J.; Ruth, T. J.; Jivan, S.; Pate, B. D. J. Fluorine Chem. 1984,25,329. Visser, G. W. M.; v. Halteren, B. W.; Herscheid, J. D. M.; Brinkman, G. A.; Hoekstra, A. J. Chem. Soc., Chem. Commun. 1984,655. Speranza, M.; Shiue, C. Y.; Wolf, A. P.; Wilbur, D. S.; Angelini, G. J. Chem. SOC.,Chem. Commun. 1984, 1448. (58) Adam, M. J.; Pate, B. D.; Newer, J. R.; Hall, L. D. Carbohydr. Res. 1983, 124, 215. Shiue, C. Y.; Wolf, A. P. J . Fluorine Chem. 1986, 31, 255. (59) Diksic, M.; Farrokhzad, S.; Colebrook, L. D. Can. J . Chem. 1986, 64, 424. Visser, G. W. M.; Boele, S.; v. Halteren, B. W.; Knops, G. H. J. N.; Herscheid, J. D. M.; Brinkman, G. A.; Hoekstra, A. J. Org. Chem. 1986,51, 1466. (60) Diksic, M.; Diraddo, P. Int. J . Appl. Rqdiat. Isot. 1985, 36, 643. (61) Chirakal, R.; Firnau, G.; Couse, J.; Garnett, E. S. Int. J . Appl. .. Radiat. Isot. 1984, 35, 651. (62) Hebel, D.; Lerman, 0.;Rozen, S. Bull. SOC.Chim. Fr. 1986,861. (63) Kosower. E. M.: Hebel. D.: Rozen, S.: Radkowski. A. E. J . O w . Chem. 1985,50,4152. (64) See, for example: Katsuhara, Y.; Des Marteau, D. J . Am. Chem. SOC.1980,102, 2681. Naumann, D.; Herberg, S. J. Fluorine Chem. 1982, 19,205. Boguslavskaya, L. S.; Ternovskoi, L. A. Zh. Org. Khim. 1983,19, 1881.
(65) Nair, K. P. R.; Hoeft, J.; Tiemann, E. Chem. Phys. Lett. 1979,60, 253.
Rozen
312
strong tendency to disproportionate to IF3 and IF5 in solution.66 Practically no reactions of IF with organic compounds have been performed. Even less is known about BrF, since this molecule has never been isolated or fully characterized despite many efforts. It was concluded that it exists at low temperature, though its complex with pyridine can be isolated.67 We have not made any attempt to isolate or purify the IF formed from the interaction of Iz with Fz in CFC13 and found that this was not necessary. The strongly polarized IF reacted quite efficiently with most types of olefins even a t -78 “C with full regio- and stereoselectivity. The trans addition is characteristic of a nucleophilic attack by the fluoride (originated a step earlier from FJ on the cyclic halonium ion.68 While IF was added across H centers as expected, BrF, similarly obtained, proved to be uncontrollable and nonselective. When, however, a proton donor such as ethanol was added to the in situ generated reagent, a considerable taming effect was observed, and clean anti addition across most double bonds took place with a full stereoselectivity and good regioselectivity. When either IF or BrF was added to enones, HF elimination could easily be induced and the Lu-iodo or -bromo enones were formed (eq 18).68 “-7%,
\\-
/WX+QF F
(18)
x
= I . Br
P I
-1%
oh’+‘
O+l
x F
X
An interesting reaction was found when iodine and bromine monofluorides were reacted with acetylenes. The first XF molecule can readily be added to form the fluoroiodo (or fluorobromo) olefin. In contrast to other methods, further addition of XF across the relatively unreactive halogenated olefin is possible. The addition is such that the most stable fluorocarbocation is initially formed, leading eventually to the corresponding gemdifluoro derivatives. Although an isolated acetylene reacts in a similar way with both IF and BrF, the latter is more active and capable of reaction also with ynones, which are not affected by IF. Even the triple bond in dimethyl acetylenedicarboxylate, which is most deactivated, is not resistant toward BrF and after 12 h at -75 “C dimethyl 2,3-dibromo-3,3-difluorosuccinate was obtained in good yield (eq 19).69 RCECR’
XF
-
RCF=CXR’
XF
While the above reaction offers a route to the important CF2moiety starting from an acetylene, the most desired transformation in this field is the conversion of a carbonyl into a CF2 group. This is usually achieved by using the unpleasant and expensive SF?O or its derivative (diethy1amino)sulfurtrifluoride (DAST).71 We have adopted a different approach to this problem, involving converting a ketone to its hydrazone, DNP, or semicarbazone derivative. The basic nitrogen atoms in these are susceptible to attack from the electrophilic iodine in IF, allowing the nucleophilic fluorine to react with the developing positive charge on the carbon, leading eventually to the desired CO CF2 transformation in reasonable yields (eq 20).72 R2CO
+
R&=NNR’2
IF
R2CFz
(20)
Conclusion Our main goal has been to demonstrate that elemental fluorine, a most neglected element in organic chemistry, is indeed a versatile reagent able to perform many selective reactions. During the 100 years since its discovery, this element has succeeded in being associated with frightening legends, which generated mythical fears and strong prejudice. Less than 20 years ago an article appeared in the Journal of Chemical Education stating that “fluorination by fluorine is unlikely to be used in normal organic ~ y n t h e s i s ” .We ~~ hope that the community of chemists will come to realize that this halogen has a great deal to offer to organic sythesis as well as toward constructing a wide variety of products with biological significance such as 2-deoxy-2-fluoro-~-glucose with or without the important 18Fradioisotope. Recently published reviews on the subject74are evidence that such a trend is indeed emerging. I a m deeply indebted to the graduate students who have brought life to the work described here. Their names are to be found i n the appropriate references. I a m also indebted to Professor D. H. R. Barton for his continuous interest i n our work and to all fluorine chemists at E. I . d u Pont de Nemours, E x perimental Station, DE, for their support and hospitality. I also thank the Fund for Basic Research of the Israel Academy of Science and Humanities and Eli Lilly & Co., I N , for partial financial support. Registry No. F,, 1182-41-4.
RCFPCX2R’
BrF
MeOOC-C=C-COOMe Me00C-CFzCBr2-COOMe
Accounts of Chemical Research
(no reaction with IF) (19)
(66) Schmeisser, M.; Sartori, P.; Naumann, D. Chem. Ber. 1970, 103, 880. (67) Naumann, D.; Lehmann, E. J . Fluorine Chem. 1975,5, 307. (68) Rozen, S.; Brand, M. J . Org. Chem. 1985, 50, 3342.
(69) Rozen, S.; Brand, M. J . Org. Chem. 1986, 51, 222. (70) BosweU, G. A.; Ripka, W. C.; Schribner, R. M.; Tullock, C. W. Org. React. 1974, 21, 1. (71) Middleton, W. J. J. Org. Chem. 1975, 40, 574. (72) Rozen. S.; Brand. M.: Zamir. D.: Hebel. D. J . Am. Chem. SOC. 1987, 109, 896. (73) Sharts, C. M. J . Chem. Educ. 1968, 45, 3. (74) Purrington, S. T.; Kagen, B. S.; Patrick, T. B. Chem. Reu. 1986, 86,997. Haas, A.; Lieb, M. Chimia 1985,39, 134. Vyplel, H.Chimia 1985, 39, 305.
